Effects of the anti-inflammatory drug budesonide on the gut microbiota and cytokine production of 13-lined ground squirrels during prehibernation fattening

Abstract

The gut microbiome is essential for maintaining organismal health. Gut microbiota may be disrupted through external factors like dietary change, which can lead to gut inflammation, resulting in obesity. Hibernating mammals develop low-grade gut inflammation when they accumulate fat deposits in preparation for hibernation, making them useful models for studying the relationship between the microbiome, inflammation, and weight gain. Nonsteroidal anti-inflammatory drugs and steroids are commonly used in humans to target gut inflammation, but how these drugs affect the gut microbiome and its stability is unclear. We investigated the effect of the glucocorticoid drug budesonide on the gut microbiome and cytokine levels of an obligate hibernator, the 13-lined ground squirrel, during the fattening season. We used 16S rRNA gene sequencing to characterize bacterial communities in the lumen and mucosa of the cecum and colon and measured proinflammatory [tumor necrosis factor-α (TNF-α)/interleukin 6 (IL-6)] and anti-inflammatory (IL-10) cytokine levels. Budesonide affected the microbiome only in the cecum lumen, where bacterial diversity was higher in the control group, and communities significantly differed between treatments. Across gut sections, Marvinbryantia and Enterococcus were significantly higher in the budesonide group, whereas Sarcina was higher in the control group. TNF-α and IL-6 levels were higher in control squirrels compared with the budesonide group, but there was no difference in IL-10 levels. Overall, budesonide treatment affected the microbial community and diversity of 13-lined ground squirrels in the cecum lumen. Our study presents another step toward developing ground squirrels as a model for studying the interaction between the microbiota and host inflammation.

NEW & NOTEWORTHY Disruptions of gut microbiota can lead to inflammation, resulting in weight gain. Inflammation can be treated with budesonide, but how budesonide affects gut microbiota is unclear. Thirteen-lined ground squirrels experience low-grade gut inflammation during prehibernation fattening, which compares with human inflammation-weight gain mechanisms. We showed that budesonide treatment decreased microbiome diversity and lead to a shift in community in the cecum lumen. Our study supports developing ground squirrels as a model for studying microbiome-inflammation interactions.

INTRODUCTION

The gut microbiome plays an essential role in maintaining homeostasis through, for example, its involvement in energy metabolism and immune system modulation, and disruption of this homeostasis can have negative effects on host health (1, 2). Shifts in the gut microbiome can be caused by external and internal factors such as changes in host diet (3), disease [reviewed in (4)], and subsequent drug therapy (5). Gut microbiome dysbiosis and reactions of the host immune system to microbial antigens may lead to gut inflammation (6). Gut inflammation can lead to destabilization of the intestinal barrier, resulting in leakage of antigens into the peripheral tissues, a condition referred to as “leaky gut,” and related immune responses (7). Leaky gut has been identified as a causative agent for metabolic conditions like insulin resistance and obesity (7). Within the microbiome, the families Prevotellaceae and Enterobacteriaceae, as well as the genera Enterobacter and Bacteroides, are more abundant in obese individuals and have been associated with increased gut inflammation (8–11). In addition, certain species within the genus Ruminococcus are associated with obesity and insulin resistance due to their role in energy harvest and fat storage (12). In contrast, higher levels of beneficial bacteria such as Akkermansia muciniphila and Bifidobacterium species have been found to decrease gut permeability, reducing the effects of microbiome dysbiosis (8, 11).

Low-grade inflammation associated with continuous antigen leaking from the gut can be mitigated using drugs such as nonsteroidal anti-inflammatories and steroids. Budesonide is a steroid in the glucocorticoid family that has a wide range of applications. For example, budesonide is used as an inhalant to reduce severity of asthma (13) and can be administered orally for gastrointestinal diseases like Crohn’s disease, ulcerative colitis, microscopic colitis, and eosinophilic esophagitis (14–16). Budesonide binds to and activates glucocorticoid receptors (GRs) in the cytoplasm of cells, allowing migration into the cell where it binds to histone deacetylase 2 (HDAC2) and CREB-binding protein (CBP). The budesonide-CBP receptor complex prevents the expression of genes that can cause inflammation. In addition, activation of HDCA2 reduces the formation of several proinflammatory interleukins and TNF cytokines. Its versatility in minimizing gut inflammation indicates that budesonide could be used as a potential treatment for other inflammatory diseases that originate in the gut, such as obesity.

Obesity is a rapidly increasing problem in developed countries (17), in part due to its association with comorbidities such as diabetes, cardiac disease, and inflammation of the gut [reviewed in (18)]. In humans, obesity-associated inflammation is primarily located in the distal small intestine (i.e., ileum) and proximal colon (19). Investigating the effect of anti-inflammatory drugs on gut inflammation in humans is challenging, as sample collection can be difficult and invasive to the host. In animal models, obesity and inflammation studies focused on the gut microbiome have largely relied on the use of laboratory rodents such as mice, which lack genetic diversity and for which a vendor effect on microbiome composition has been reported (20).

Mammalian hibernators serve as a natural model of obesity because they accumulate large fat deposits in late summer and fall to survive the long period of prolonged fasting and inactivity that characterizes hibernation. The increased adiposity is believed to result from hyperphagia, hormonal changes, and decreased metabolism (21, 22). Thirteen-lined ground squirrels (Ictidomys tridecemlineatus) are obligate hibernators, and during the active season, they generally increase their caloric intake to develop high-energy fat stores (23). During this period of rapid fattening, they develop insulin resistance and inflammation of the gut, white adipose tissue, and muscles (23). Similar patterns of metabolic inflammation and insulin resistance are also detected in obese humans, which suggests that fattening ground squirrels could be used as a model for rapid adiposity.

Drug therapies can affect the gut microbiota and lead to dysbiosis, which can hinder recovery (5). As obesity already has numerous comorbidities and has a direct connection to the microbiota, additional negative effects from drug therapy are undesirable. The objective of our study was to investigate the effect of the anti-inflammatory drug budesonide on the gut microbiome community and gut cytokine levels of 13-lined ground squirrels during active season fattening. We hypothesize that 1) budesonide treatment will suppress gut inflammation in squirrels, resulting in a stable gut microbiome over the duration of our experiment, whereas control group microbiomes will naturally shift over time in response to metabolic inflammation associated with fattening and 2) levels of proinflammatory cytokines will be depressed in response to the budesonide treatment.

Our study is a counterpart to a study by Zur Tulod et al. (24), who investigated the effect of mesalazine on metabolic inflammation and seasonal fattening in 13-lined ground squirrels. They showed that mesalazine was effective at reducing metabolic inflammation in fattening ground squirrels, but surprisingly they did not detect an effect of treatment on the cecum and colon microbiome. However, budesonide was shown to be more effective at reducing Crohn’s disease inflammation than mesalazine, with 62% versus 36% (25), which may result in a detectable effect on the microbiome. As of yet, budesonide has not been tested as an anti-inflammatory treatment for this obesity comorbidity. Since the microbiome is an integral part of gut health and the initial chain in the path to metabolic inflammation, we focused on the effect of budesonide on the gut microbiome.

METHODS

Study System

Gravid female 13-lined ground squirrels were trapped in Wisconsin in spring 2019, gave birth in captivity at the University of Wisconsin, Oshkosh, and their pups hibernated in captivity over the 2019/2020 winter. Following successful hibernation in captivity, 40 now yearling squirrels were used for our study. Squirrels were housed in cages (25 cm width × 48 cm length × 20 cm height) and fed a base diet [ad libitum Teklad Global 18% Protein Diet (Envigo, Madison, WI) plus ∼6.5 g of sunflower seeds weekly] and water ad libitum for 7 wk, after which they were randomly assigned to the control (n = 20; 9 males and 11 females) or budesonide (n = 20; 9 males and 11 females) experimental groups. Light/dark cycles were adjusted weekly to mimic natural sunrise and sunset in Oshkosh, WI, and body mass was recorded weekly throughout the study. All protocol and procedures for this study were approved and overseen by the University of Wisconsin Oshkosh Institutional Animal Care and Use Committee (Protocol No. 0026-000298).

At 8 wk posthibernation, squirrels in the experimental group were fed a diet containing ∼0.015 mg/g budesonide, equivalent to 1 mg/kg dose (Envigo, TD.200275, Madison, WI) for 10 wk (until 18 wk posthibernation). Control squirrels remained on the base diet for the same time period. During the experiment, a subset of squirrels was euthanized in weeks 5 (n = 10/group) and 10 (n = 9 or 10/group) of the experimental period. One male squirrel died before week 10 of causes unrelated to the study. Each squirrel was anesthetized with 5% isoflurane in an induction chamber, weighed, and decapitated.

We collected lumen and mucosa from the small intestine, cecum, and colon. Colon and small intestine lumen contents were collected by removing each and gently squeezing the lumen material into a sterile Petri dish. The cecum was removed to a sterile Petri dish, cut open with sterile instruments, and a sterile spatula was used to gently remove the contents. Subsequently, the tissues were cut open, gently rinsed with sterile saline to remove excess lumen contents, and mucosa was collected by scraping the inside of the tissue with a sterile glass slide. Samples were transferred to 2-mL cryotubes and snap-frozen in liquid nitrogen. Portions of the small intestine and proximal colon tissues were flash frozen in liquid nitrogen for immunological analyses. All collected tissues were stored at −80°C until further analysis.

Cytokine Measurements

We used enzyme-linked immunosorbent assays (ELISAs) to measure cytokine concentrations in tissues. Tissue pieces were weighed out for each ELISA (∼50–75 mg) and homogenized in PBS containing protease inhibitor cocktail. Samples were mechanically homogenized using the Tissue Master 125 (OMNI International, Lake Villa, IL), homogenates were centrifuged for 20 min at 10,000 rpm at 4°C, and the supernatant was used in the ELISA. Rat-specific ELISA kits were obtained for tumor necrosis factor (TNF)-α, interleukin (IL)-6, and IL-10 from BD Biosciences (San Jose, CA). Sequence alignment showed >98% homology between rat and ground squirrels for TNF-α (>99%), IL-6 (>99%), and IL-10 (>98%) (23). ELISA results were normalized to protein concentration obtained from bicinchoninic acid (BCA) protein assays of the same supernatants (Pierce Biotechnology, Waltham, MA).

DNA Extraction and Sequencing

We extracted DNA from the small intestine, cecum, and proximal colon contents (mucosa and lumen, ∼0.15 g/extraction) using Qiagen RNeasy PowerMicrobiome Kits (Qiagen, Hilden, Germany) according to manufacturer’s protocols with omission of the DNase digestion step. DNA concentration and purity were measured with a Qubit Fluorometer (Invitrogen, Carlsbad, CA), and extractions were stored at −80°C. The V4 region of the 16S rRNA gene was PCR amplified in triplicate using universal bacterial primers 515 F (5′-GTGYCAGCMGCCGCGGTAA-3′) and 806 R (5′-GGACTACNVGGGTWTCTAAT-3′) with attached 12-bp Golay barcodes following to the Earth Microbiome Project protocol [available in the public domain at www.earthmicrobiome.org (26)]. Amplification success and purity were checked by gel electrophoresis. We pooled PCR products and cleaned DNA using the AxyPrep Mag PCR cleanup kit (Axygen, Union City, CA). Final DNA concentrations were determined using the Kapa Library Quantification Kit (Roche Sequencing Solutions Inc., Pleasanton, CA), and samples were combined in equimolar amounts. Libraries were sequenced paired-end (300 × 2 bp) using Illumina MiSeq platform (MiSeq v2 Reagent Kit). Sequences were uploaded to NCBI database under BioProject PRJNA1040919 (Accession No. SRX22567264). R code is available at 10.6084/m9.figshare.26308363.

Sequence Processing and Microbiota Analysis

All sequence and statistical analyses were conducted in R (27). Sequences were analyzed using the DADA2 pipeline in R with default parameters (28). In short, sequences were trimmed and filtered, chimeras were removed, and the Silva database (v. 138.1) was used for taxonomic classification to genus level. Potential contaminants were removed using the decontam package (v. 1.16.0). Sequences identified as chloroplast and mitochondrial sequences were removed from the dataset, as well as sequences that did not align to bacteria. The DECIPHER package in R (29) performed a multiple alignment, and the phangorn package v. 2.4.0 (30) constructed a phylogenetic tree of the microbial amplicon sequence variants (ASVs).

Statistical Analysis

Samples were rarified to 13,184 sequences/sample, which was the lowest number of sequences detected in a sample with a 10,000 read minimum. All samples with lower read numbers had fewer than 134 reads/sample. To compare alpha diversity among sample variables, we calculated the observed number of ASVs (Observed) and Shannon’s diversity index (Shannon’s H). The observed number of ASVs only represents the number of ASVs detected in the sample (also referred to as richness), whereas Shannon’s diversity index represents a combination of the number of ASVs and the evenness of their distribution. Analysis of variance (ANOVA) was used to assess the differences in diversity in our variables. We calculated pairwise differences within variables using Tukey’s honestly significant difference (HSD) tests with a Bonferroni correction for multiple comparisons. We analyzed microbiome characteristics using the phyloseq package (v. 1.40.0) (31) and visualized results using the ggplot2 package (v. 3.36) (32). We used beta diversity to provide a measure of how microbiome samples differed from each other in microbial composition and structure. To visualize beta diversity, we applied principal coordinate analyses (PCoA) using the Bray–Curtis distance matrix (33). To determine which variables contributed to the most variation in microbiome composition, permutational multivariate analysis of variance (PERMANOVA) was used with the adonis2 function from the vegan package (v. 2.6-2) (34). In addition to our variables, we correlated three immune measures (TNF-α, IL-6, and IL-10) collected in this study to gut microbial diversity community metrics.

We investigated the associations between bacterial taxa and metadata (treatment, gut section, and sampling week) using the microbiome multivariable associations with linear models (MaAsLin2) package in R (35). We normalized [trimmed mean of M-values (TMM) normalized] counts within the model and used a minimum prevalence rate of 0.4 (taxon needs to be present in at least 40% of the samples). False discovery rate (FDR) corrections were applied to P values to correct for multiple comparisons.

Weight gain was compared between squirrels on the budesonide and control treatments using a linear mixed-effect model with squirrel ID as a random factor to account for the lack of independence between measurements. Mixed-effect models were compared using ANOVAs. To assess the effect of budesonide on cytokine levels in the colon, we compared TNF-α, IL-6, and IL-10 levels in control and treatment squirrels at week 5 and 10 of sampling using ANOVAs.

RESULTS

There was no significant difference between the rate at which squirrels in the control and treatment groups gained weight (ANOVA, F_1,64 = 1.037, P = 0.361), although squirrels receiving the budesonide treatment tended to be heavier throughout the experimental period (Fig. 1). Sex did not affect the rate of weight gain in squirrels (ANOVA, F_1,64 = 0.103, P = 0.749).

Figure 1.

Absolute body weight (A) and change in body weight percentage (B) of 13-lined ground squirrels receiving different treatments from 0–18 wk posthibernation. The shaded area represents the experimental period. Significance (ANOVA) was set at alpha = 0.05 and no significance is depicted as not significant (N.S.).

Microbiome Diversity and Composition

We were unable to extract DNA in sufficient concentrations for sequencing from the small intestine samples and could therefore not examine the microbiome in this gut section. We successfully sequenced 138 samples from the cecum and colon lumen and mucosa. Rarefaction resulted in the loss of 22 samples. Sequencing was unsuccessful in 7 colon mucosa samples and 15 cecum samples (13 mucosa and 2 lumen), despite repeated sequencing efforts.

Alpha diversity.

Alpha diversity only significantly differed between treatments in the cecum lumen (Fig. 2; ANOVA, F_2,41 = 15.18, P < 0.001). The budesonide treatment showed a lower alpha diversity than the control group at both sampling points in the cecum lumen (week 5: Tukey’s HSD, P = 0.045; week 10: Tukey’ HSD, P < 0.001), but no within-treatment differences were observed between the sampling times. Observed number of ASVs did not differ among treatments or sampling times (data not shown). We did not detect an effect of sex on either measure of alpha diversity.

Figure 2.

Microbiome Shannon’s alpha diversity at 5 and 10 wk of treatment from different gut sections harvested from 13-lined ground squirrels. Significance was set at alpha = 0.05, and no significance is depicted as not significant (N.S.). Pairwise significance [Tukey’s honestly significant difference (HSD)] is shown using letters, where different letters represent significant pairwise differences. Control samples from the colon lumen did not sequence successfully, so only budesonide samples were shown earlier.

Beta diversity.

From the colon lumen, only samples from the treatment group from sampling week 10 sequenced successfully and we therefore omitted this group from our statistical analyses. Treatment significantly impacted gut microbiome community composition across all sections combined (Fig. 3; PERMANOVA, R² = 0.054, P < 0.001). Although not significant, we detected a weak effect of sampling week on the overall gut microbiome composition, but with a low contribution of the total microbiome variation of 1.2% (PERMANOVA, R² = 0.012, P = 0.053). Treatment explained 23.4% of the variation in the microbiome communities in the cecum lumen (PERMANOVA, R² = 0.234, P < 0.001), but did not significantly impact the cecum and colon mucosae (colon: P = 0.743, cecum: P = 0.540). Sampling week did not significantly affect beta diversity in any gut sections for either treatment (P = 0.091–0.868), and we found no significant difference in beta diversity between sexes (PERMANOVA, P = 0.091–0.317).

Figure 3.

Principal coordinates analysis for 3 different tissues collected from 13-lined ground squirrels subjected to different treatments and sampled at 2 time points (5 wk and 10 wk). Colon lumen results were not shown due to poor sequencing success of control samples. Significance [permutational multivariate analysis of variance (PERMANOVA)] was set at alpha = 0.05, and no significance is depicted as not significant (N.S.).

Community composition.

On the phylum level, a majority of the gut microbiomes of squirrels consisted of Firmicutes (control: 41.8%; budesonide: 43.2%), followed by Bacteroidota (control: 7.1%; budesonide 5.9%), and Verrucomicrobiota (control: 3.2%; budesonide: 2.5%) (Fig. 4A). Bacilli and Clostridia dominated the microbiome of both control and budesonide squirrels on a class level, with higher relative abundance of Bacilli in budesonide squirrels compared with controls (Fig. 4B). On the genus level, squirrels on the budesonide treatment had a higher relative abundance of Enterococcus and a lower abundance of the Christensenellaceae R-7 group (Fig. 4C). Marvinbryantia (family: Lachnospiraceae) was the fourth most abundant genus in squirrels on the budesonide treatment but was not among the top 10 genera in the control group of squirrels.

Figure 4.

A–C: relative abundances of phyla, classes, and the top 10 most abundant genera in samples collected from squirrels at week 5 and 10 of their respective treatments. D–F: genera that were significantly associated with either the control or budesonide treatment. Read counts were trimmed mean of M-values (TMM) normalized.

Taxa associations.

We detected two genera that were significantly associated with the budesonide microbiome (Fig. 4, E and D; Marvinbryantia P_FDR = 0.003, Enterococcus P_FDR = 0.042) and one species that was associated with the control microbiome (Sarcina P_FDR = 0.042) (Fig. 4D). We did not find any genera significantly different among gut sections or sampling weeks.

Cytokines

There were no differences in any cytokines with budesonide treatment in the small intestine (data not shown), but there were effects in the colon. IL-6 was significantly higher in the colon of squirrels on the budesonide treatment compared with the controls (Tukey’s HSD, P = 0.003) at week 10, but not at week 5 (Fig. 5A). In addition, IL-6 levels were significantly higher in the colon of control squirrels sampled in week 5 compared with that of control squirrels sampled in week 10 (Tukey’s HSD, P = 0.04). IL-10 levels were significantly lower in week 10 than in week 5 (ANOVA, F_1,35 = 4.141, P = 0.05), but no effect of treatment was observed (Fig. 5B; ANOVA, F_1,35 = 1.575, P = 0.218). TNF-α levels in the colon were lower in budesonide-treated squirrels compared with control squirrels (Tukey’s HSD, P = 0.02) at 5 wk; however, there were no differences in TNF-α levels at 10 wk (Fig. 5C). In general, males responded better to budesonide treatment than females. Females had significantly higher TNF-α independence of treatment (ANOVA: F = 11.69, P = 0.016), and IL-10 was not affected by budesonide at the 5-wk time point in females, but was in males (ANOVA: F = 11.92, P = 0.012) (Supplemental Fig. S1). No sex difference was detected for IL-6.

Figure 5.

IL-6 (A), IL-10 (B) and TNF-α (C) levels in the colon of experimental and control squirrels at 5 and 10 wk of their respective treatment. Numbers represent significant P values (ANOVA).

DISCUSSION

The objective of our study was to investigate the effect of the anti-inflammatory drug budesonide on the gut microbial community and gut cytokine levels of 13-lined ground squirrels during active season fattening. We detected an effect of budesonide on the microbiome diversity and community composition, but significant differences were only detected in the cecum lumen.

We found significant differences in alpha diversity and microbial community composition between our budesonide and control group in the cecum lumen, but not in the cecum mucosa and the colon lumen. The alpha diversity of squirrels receiving the budesonide treatment was lower in both week 5 and 10, compared with the control group. Given that lower alpha diversity is often associated with increased gut inflammation (36, 37), if budesonide had indeed decreased inflammation as expected for an anti-inflammatory drug, we should have seen a higher alpha diversity in the treatment group. Zur Tulod et al. (24) found a decrease in alpha diversity with time in the colon overall, but no effect of their anti-inflammatory drug mesalazine. However, we did not find a change in alpha diversity in the colon mucosa and were unable to make the temporal comparison in the colon lumen due to low sequencing success in that section. Similar to Zur Tulod et al. (24), we detected lower alpha diversity and a different community composition in squirrels in the treatment group compared with controls in the cecum lumen, but not in other sections (24). Budesonide has been shown to reduce gut inflammation in a wide range of inflammatory bowel diseases in humans and rodents (16, 38, 39). We only found a treatment effect on diversity and composition of microbiota in the cecum lumen, which suggests that budesonide may affect the microbiome locally, with little impact on the squirrel as a whole. In addition, it is possible that budesonide affects the microbiome in a way that is unrelated to the level of inflammation, but the mechanism behind this is unclear. We hypothesized that budesonide treatment would suppress late-season gut inflammation in squirrels, resulting in a stable gut microbiome community over time. We did not find an effect of sampling week on the microbial community composition, or beta diversity, in the budesonide treatment, which supports our hypothesis. However, we also failed to see a shift in the microbiome community in our control group and therefore cannot attribute the microbiome stability to the potentially reduced inflammation caused by budesonide supplementation.

Upon ending hibernation, the squirrel microbiome undergoes drastic changes in diversity and composition due to changes in host behavior and physiology, in particular, the reintroduction of food (40, 41). Squirrels in our study had been on a regular diet for 7 wk posthibernation before they were switched to the experimental diet. It is possible that by week 8 squirrels had already reached a stable community that was little affected by fattening in the following 5 wk. Similar conclusions were drawn by Hatton et al. (42) in arctic ground squirrels fed diets differing in fat and calorie content (42). Coincidentally, weight gain in our squirrels was minimal between weeks 5 and 10 in both budesonide and control groups, which may be the reason that we do not see a shift in microbiome associated with fattening. Phylogenetic diversity and numbers of unique operational taxonomic units (OTUs) are lowest in late winter and highest 2 wk after refeeding in spring for 13-lined ground squirrels (40); however, no samples were collected later in the active season. In arctic ground squirrels, Stevenson et al. (43) did not find evidence of shifts in the gut microbiota associated with fat deposition during the active season. Although we detected a similar pattern, host species and sampling age were different in their study.

We detected two genera that were significantly associated with the budesonide microbiome, Marvinbryantia and Enterococcus. The Marvinbryantia genus only contains one known species, Marvinbryantia formatexigens, which has previously been isolated from human feces (44). M. formatexigens is an anaerobic acetogen that converts H₂ into the short-chain fatty acid (SCFA) acetate. They play a role in energy metabolism and were found to consume oligosaccharides and increase the succinate yield when transplanted into gnotobiotic mice (44). Few studies exist on M. formatexigens, and their role in the gut microbiome and gut inflammation is not well understood. However, acetate, a product of their metabolism, is positively associated with epithelial tissue maintenance and wound healing (45), has been shown to lower colon inflammation in mouse models (46) and may be involved in limiting the effects of enteropathogens (47). Weight loss associated with a low nutrition diet was associated with a decrease in Marvinbryantia in goats, while a standard nutrition diet resulted in higher Marvinbryantia abundances as well as higher acetate levels (48). Similarly, increased abundances of Marvinbryantia were detected in overweight children compared with those of normal weight (49), but the opposite trend was observed in adults with obesity (50). Moving forward, investigating the relationship between anti-inflammatories, weight gain, and microbial SCFA production using functional metagenomics methods could be an interesting target.

Enterococcus species were significantly more abundant in microbiotas from squirrels on the budesonide treatment compared with the control. The Enterococcus genus comprises over 50 species and contains a number of common commensal bacteria found in the mammalian gut, most notably Enterococcus faecalis and Enterococcus faecium. Administration of the E. faecium SF68 strain to mice with diet-induced obesity caused their gut microbiome to shift toward a community associated with reduced inflammation and higher SFCA production (51), which suggests that budesonide supplementation could have additional indirect anti-inflammatory effects through modulating of the microbiome. Unfortunately, the classification depth of 16S rRNA gene sequencing does not allow for species level identification, thereby limiting our interpretation of the involvement of this large genus.

Due to low DNA concentrations, we were unable to investigate the microbiota in the small intestine and its response to budesonide. In a previous study, we showed that the small intestine of 13-lined ground squirrels contained a low-diversity microbiome and a distinct microbial community compared with the cecum (52). The small intestine is located upstream of the cecum and predominantly involved in enzyme production and absorption of nutrients. One of the uses of budesonide is to treat Crohn’s disease in the small intestine, and, depending on the delivery method, budesonide absorption rate can be high in this gut section (53). It is possible that budesonide has a different effect on the distinct small intestine microbiome, but we unfortunately were not able to confirm this.

Budesonide treatment elicited different responses in proinflammatory cytokine production by the host. We detected significantly higher levels of the proinflammatory cytokine, IL-6, in the budesonide treatment group later in the season. Budesonide inhibited IL-6 bioactivity in mouse cell lines (54), but the inhibitory effect was highly dose dependent. It is possible that the calculated budesonide dosage for our squirrels was insufficient for IL-6 inhibition if budesonide was less effective or metabolized faster in squirrels than humans. However, at present we cannot explain why we found higher IL-6 levels associated with budesonide treatment than the control treatment.

Levels of the proinflammatory cytokine TNF-α were significantly lower early in the season in our budesonide treatment compared with the control, indicating a possible anti-inflammatory effect. We initially detected lower TNF-α levels in budesonide treatment at sampling week 5, but no difference was observed later in the season. Budesonide administration was constant throughout the experimental period, and it is possible that the amount of budesonide used was not sufficient to mitigate the increase in metabolic inflammation occurring late in the active season. It is unclear whether the effects of budesonide “wear off” after a time or if this is a result of the animals entering their prehibernation state, which has been shown to be a more proinflammatory environment (23). With regard to cytokine levels, budesonide treatment showed similar results as administration of another anti-inflammatory drug, mesalazine (24), which supports the potential application of budesonide for metabolic inflammation.

One of the cytokines we tested in our study, IL-10, is an anti-inflammatory cytokine that reduces the production of proinflammatory cytokines TNF-α and IL-6 and thereby attenuates the inflammation associated with these cytokines (55). We detected a significant decrease in IL-10 levels later in the season in our control group, a result that was also recorded by Sonsalla et al. (23) in the same host species. This decrease appeared to be independent of the presence of inflammation, as we also detected lower IL-10 levels in squirrels receiving budesonide treatment. Similar results were found with administration of the anti-inflammatory drug mesalazine, albeit only early in the season (24).

Overall, we detected some effects of budesonide treatment on the gut microbiota community and diversity of 13-lined ground squirrels, albeit localized to the cecum lumen. If and how these changes in the microbiome affect the host would require a functional approach, which we were not able to do with our sequencing method. How anti-inflammatory drugs influence the microbiota presents an interesting research avenue, especially with respect to gastrointestinal inflammation, as the change in the microbiota is often responsible for said inflammation. In addition, our study presents another step toward developing ground squirrels as a model for studying the interaction between the microbiota and host inflammation with the potential to generate results that are translatable to humans.

DATA AVAILABILITY

Data will be made available upon reasonable request.

SUPPLEMENTAL MATERIAL

Supplemental Fig. S1: https://doi.org/10.6084/m9.figshare.26761759.v1.

GRANTS

This work was supported by National Institute of General Medical Sciences Grants R15GM124586 (to C.C.K.) and P20GM130443 (to K.G.).

DISCLAIMERS

The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health.

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the authors.

AUTHOR CONTRIBUTIONS

C.C.K. and K.N.D. conceived and designed research; J.Z.T. and C.C.K. performed experiments; K.G. and J.Z.T. analyzed data; K.G., J.Z.T., C.C.K., and K.N.D. interpreted results of experiments; K.G. prepared figures; K.G. drafted manuscript; K.G., J.Z., C.C.K., and K.N.D. edited and revised manuscript; K.G., J.Z.T., C.C.K., and K.N.D. approved final version of manuscript.